Een kalibratiefile is eventueel mogelijk als er genoeg interesse is, dan doe ik de hele batch in 1x. Wel ten opzichte van een iSEMcon EMX7150, en niet ten opzichte van een "echte" lab grade mic zoals iets van B&K.
I made a very compact, high performance but low cost phantom powered microphone for performing accurate low frequency in box measurements of an assembled ported box.
We all know it's a beech to get accurate measurements of a ported speaker in your mancave right?
You either need a large empty space for a groundplane measurement, or you have to scale and sum seperate measurements of the woofer and port, but then you still wonder if it is accurate.
Another method that IS easy and accurate is to measure inside of the box and apply a 12dB/oct tilt to the measurement. Signal to noise will be high, and the result should be similar to what you get (at LF) with a good groundplane measurement or in an anechoic chamber.
With larger subs you could drop the mic trough the port, but how do you get that microphone through a smaller L shaped port or a port that is infront of a cabinet wall? And what about it rattling against something? You'll have to disassemle the box to place the mic trough the woofer opening or you'll have to drill a hole to stick the microphone trough.
I figured that making a very small microphone would solve most of the issues with an in box measurement. The microphone can be fished trough almost any port, and by using a flexible cable you can let the microphone hang past the port inside of the box. A lavelier microphone type windscreen can protect the microphone from rattling against anything if it cannot hang free.
Last month I learned about simple and compact low noise linear power supplies, this was the last puzzle piece needed to pull off this idea with high performance and low cost, so I thought lets give it a go!
I finished the first couple of boards yesterday and did some testing and measuring, as well as a test speaker measurement. See the images for the details results.
Features:
Very compact
- Currently the microphone is directly soldered to the cable and it measures just 3*10mm. The microphone PCB that will be used eventually is in production with the PCB fab and it will be 3*25mm. All of the electronics are on a PCB housed inside of the XLR plug.
- The cable is from the Mogami Ultraflexible series. It was developed by Mogami to handle continuous rapid movement. The cable is just 2,7mm in diameter.
- The design uses one of the highest performance MEMS microphones of those that are currently available. Arguably the best when it comes to overall performance.
- Most microphones we use as speaker builders are specified as omnidirectional, but this is never completely true.
The larger the microphone front, the lower the frequency to where it is more sensitive to sound direcectly infront of it. You can see it as reverse baffle step.
Only with an 1/8 inch (3,175mm) or smaller microphone you'll get near actual omnidirectivity up to 20kHz.
Based on dimensions this microphone should theoretically see a difference in sensitivity between sound from front and back of 0,5dB to 1,5dB at 20kHz.
- Runs off of 48V phantom power. Might work at 24V too, but I haven't tested it yet. Most interfaces are 48V anyway.
- The noise measured with a calibrated SPL meter is 29dB(A), but I think it could be lower if I improve the shielding and if I had a more quiet environment.
In the acoustic FFT (512k, 32AVG) noise measurements I was surprised to see noise levels below 0dB SPL above 100Hz.
- The sensitivity is 16,03mV/Pa (-36dBV/94dB). Relatively speaking this is quite high, this can be beneficial for cheap soundcards that have noisy inputs. This sensitivity results in an 1Vrms (0dBV) output signal level with 130dB acoustic input at the microphone.
- With a noisefloor of below 30dB and 10% distortion SPL of 134dB, the dynamic range is at least 104dB.
- The MEMS microphone elements are matched to +/-0,5dB by the manufacturer.
- The frequency response is within +/-1dB from 20kHz to 15kHz, with the +/-3dB points at around 10Hz and 20kHz (relative to 1kHz). This is within IEC 61672-1 Class 1 tolerance.
The two measured MEMS microphones that I tested were randomly picked, I could do "binning" to keep only the microphones with the least frequency response deviation.
I think the MEMS microphones are more flat than measured due to errors/deviation. The microphone used as reference (iSEMcon EMX-7150) was calibrated against a Bruel & Kjaer 4133 by iSEMcon, but their method, stability and error is unknown. Another source of error could likely be my measurement method with the two mics very near eachother, the reference microphone front is forming a baffle for the MEMS microphone. I should space them slightly, will be continued.
- To keep cost down everything is kept as simple and effective as possible. Just a custom PCB with mostly discrete components except a dual channel opamp and the MEMS microphone itsself. No custom metal parts, the microphone top of the PCB will get a thin layer of epoxy for protection instead.